The long-term goals of CA100019 are to improve the outcome of unrelated donor (URD) hematopoietic cell transplantation (HCT) and broaden the application of mismatched HCT through a better understanding of the immunogenetics of the transplantation barrier. We hypothesize that undetected variation within the gene- dense MHC could contribute to risks after HLA-matched and HLA-mismatched URD HCT. We developed a novel long-range phasing method to link HLA alleles in unrelated individuals and demonstrated that haplotype mismatching is associated with higher risk of clinically severe acute graft-versus-host disease (GVHD) after HLA matched and HLA mismatched HCT. Patients mismatched for both HLA alleles and haplotypes had significantly lower survival. Haplotype-mismatched donors and recipients had a greater degree of disparity for MHC-resident variation compared to haplotype-matched pairs, and undetected disparity was associated with increased post-transplant risks. These data demonstrate that the haplotype is a proxy for transplant risk. To define the risks of GVHD, relapse, TRM and mortality in a larger independent clinical transplant population, we will refine our long-range phasing method and develop short-range phasing methods to map variation (Specific Aim 1). We will determine the extent to which haplotype matching can lower risks of GVHD and TRM, while preserving graft-versus-leukemia effects (Specific Aim 2). We will identify novel MHC-resident variation associated with transplant risks (Specific Aim 3). CA100019 will provide a powerful practical immunogenetic approach for optimizing URD HCT and for broadening the use of HLA mismatched URDs. Finally, CA100019 will contribute novel genetic data needed for hypothesis-driven and exploratory gene mapping. PUBLIC HEALTH RELEVANCE: We developed a tool for studying human chromosomes and genes, and discovered ways to improve the results of unrelated transplantation for the treatment of blood disorders. We will use the tool to find the genes that are important to transplant success. [unreadable] [unreadable] [unreadable]